Estimating Probability of Failure of Composite Laminated Panel with Multiple Failure Modes

Abstract

As composite structures play a key role in aircraft structure design, assessing probability of failure (PF) of the composite structures has become important to ensuring safety of the aircraft design. To assess the PF of the design, computational models are frequently used to estimate probability of the composite structures but the computational models have uncertainty in predicting failures due to lacking knowledge of composite structure failures. Structural tests can provide additional knowledge of the failures to effectively reduce the uncertainty in the computational model. The presence of multiple failure modes due to variabilities is the other concern because correlated failure modes causes difficulty to estimate PF and it brings significant error for use of the effective reliability analysis methods, such as FORM or SORM. Undetected failure mode is the other issue relevant to the multiple failure modes. A single test can detect one failure mode, each failure modes has different probability of occurrence. Hence, a existing failure mode may not be detected and neglected. Possibly we have significantly under estimated PF. In the study, we make an uncertainty quantification model to estimate PF for correlated failure modes. Also the probability of failure occurrence of each failure modes is also estimated with the model. The model is demonstrated through a simple example and PF estimation of a curved composite panel with a hole is made. From the numerical study, it is observed that performing a single failure test can significantly reduce conservativeness on estimating PF. Test for a given probability of failure allows lighter design.